AARON JUDY, GISP Esri Certified Enterprise Geodatabase Management Professional GIS Certification Institute (GISCI) GIS Professional Certification Certified Scrum Master Ninja (No, really!)
MARICOPA COUNTY ASSESSOR S OFFICE Fourth Largest Assessment Jurisdiction in the Nation 60,000 50,000 40,000 Ownership Changes (sales) 271,260 Parcel Splits (new creates) 2,000 New Subdivision creates 345 PERMITS WORKED Land Mobile Homes Vacant Land 167,902 11% Commercial 70,966 5% Apartment 29,261 2% Mobile Unit 36,590 2% Agriculture 15,025 1% Single Family Residential/Condo 1,228,066 79% 30,000 Commercial 20,000 10,000-2007 2008 2009 2010 2011 2012 2013 Residential 2014 PARCELS BY TYPE
SO WHAT IS ATLAS ANYWAY? Automated Version Reconcile and Post Version Reconcile Order Prioritization Based on Historic Data Automated Nightly Compression Reconciliation Resilience Failed Reconcile Management Automated Self-Recovery From Failed Compression Automated Version Cleanup User Feedback \_(ツ)_/
EDITING ENVIRONMENT (PRE-ATLAS) PostgreSQL Oracle Spatial Custom Built Parcel Management System Esri Parcel Editor (Spatial Data) Esri Workflow Manager Area of Interest Feature Class Esri Data Reviewer
OBSTACLES Reconcile and post can be resource and time intensive Some users create larger, more complex versions Lack of metrics or data for comparison Limited hours in a day Nightly ETL processes DB locks Non-database savvy users handling reconcile and post Locked versions Complex geodatabase administration functions/concepts Deciding the best order to reconcile and post We process versions differently
But, what s wrong with esri s suggested reconcile order? NOTHING IS WRONG JUST A DIFFERENT APPROACH
CONCERNS Users purposely crashing ArcMap to stop reconcile It was taking too long I had to catch my bus I thought it would block the compress Interference with nightly compression Administrative connections as SDE user Mapping group was about to purchase additional machines to supplement the machines that were busy reconciling and posting User complaints regarding reconcile and post performance
GROUND WORK Teach the Machine to Decide Business rules + Human intuition Log Collection Actual metrics + Build history Bridge Systems Link databases Maintain data models Maintain support
BRIDGING SYSTEMS PostgreSQL SQL Server Oracle Spatial Custom Built Parcel Management System Esri Parcel Editor (Spatial Data) ODBC Link Database Esri Workflow Manager Area of Interest Area of Interest Feature Class Esri Data Reviewer Utilize SQL Server s ODBC DB Linking
TEACHING THE MACHINE THE CREATION OF ATLAS GOALS Prioritize jobs (versions) intelligently Asynchronous Ensure full nightly GDB compression Gracefully recover from user mistakes Maintain geodatabase security Minimize costs Increase efficiencies Decrease interruptions Remove human processes LESSONS If you can say it, you can code it The only reason for time is so that everything doesn't happen at once. Albert Einstein ArcSDE locking, just go with it, it s there to protect you Don t bypass ArcObjects! Successful doesn t always mean successful, capture all messages Design resilience into the process Plan to monitor in advance If you aren t sure, log it
PRIORITIZATION TEACHING THE MACHINE INTUITION Age Oldest Area Size Biggest 7 PM Day 1 12 Noon Day 2 Utilize the larger time window for potentially longer reconciliation Version State Size Biggest Utilize the larger time window for potentially longer reconciliation Crunch Time 12 Noon 6 PM Area Size Smallest Less likely to interfere with nightly compression Potential for a greater number of smaller versions to be completed Version State Size Smallest Less likely to interfere with nightly compression Potential for a greater number of smaller versions to be completed Percent of User s Versions Over Threshold Smallest Less likely to interfere with nightly compression Potential for a greater number of smaller versions to be completed Percent of User s Volume Overall Largest Presumes user capability and trust based on workload Percent of User s Jobs R&P Time Overall Smallest Users whose versions take less processing time overall present potential for more a greater number of smaller versions to be completed Age Oldest Increases constraints each hour until compress
STACK Prioritization Services Intuition meets workload Atlas Core Processes Versions via ArcObjects Reporting Services Building machine experience Staging Services Monitor Interface Failed job review/resubmission Status override Prioritization Services Version Management Based on Business Needs Scheduling & Proven History Reporting Services Logging History Error Reporting Staging Services Testing Monitor Interface Human Feedback
SYSTEM FEEDBACK USER LEVEL RSS ADMINISTRATIVE LEVEL Kiosk Monitor
TECHNOLOGIES BEHIND ATLAS CORE MONITOR Python Arbitrary PHP Web Services (not RESTful) Windows Scripting SQL HTML5 CSS3 PHP JavaScript
MONITOR INTERFACE DEFAULT
MONITOR INTERFACE WORKING
MONITOR INTERFACE WARNING
MONITOR INTERFACE CRITICAL
ATLAS KIOSK
RETURN ON INVESTMENT Cost (-) Developer Time = $1,800 Plus One(1) Cube of Mt. Dew and Five(5) Energy Drinks = $17.42 Savings (+) New PCs and Support = $3,500 GIS User Labor Per Year = $23,000 First Year Savings $26,500 - $1,800 = $24,700 Savings Each Year After = $23,375 Five(5) Year Savings Projected = $118,200 Version Throughput Has Increased From 7-8 to 15-18 Per Day (+) Monetized Productivity Improvement Per Year = $11,500 New First Year Savings $24,700 + $11,500 = $36,200 New Savings Each Year After $23,375 + $11,500 = $34,875 New Five(5) Year Savings Projected = $175,700
QUESTIONS
Esri FIN CREDITS Resources Tutorials Components www.esri.com developers.arcgis.com/javascript/ Clouds www.clicktorelease.com/blog/howto-make-clouds-with-css-3d# All logos, names, & titles are property of their respective owners. Charts www.fusioncharts.com